health-audit
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external, untrusted codebase data to generate health reports.
- Ingestion points: The agent processes source code, schemas, and configurations from the repository it is auditing to fulfill the mission defined in
SKILL.md. - Boundary markers: The skill establishes a "NON-NEGOTIABLE" Evidence Rule and a strict "Validation Phase" (Phase 3). These instructions force the agent to corroborate every finding with concrete code evidence and file paths, effectively preventing the agent from blindly following instructions that might be embedded in the audited code.
- Capability inventory: The skill primarily utilizes file-reading capabilities to perform inspections across architecture, security, and maintainability categories.
- Sanitization: The skill relies on rigorous logical verification and evidence tiers (Verified, Probable, etc.) to ensure its output is driven by structural analysis rather than external prompts hidden in comments or documentation within the codebase.
Audit Metadata